機械外文文獻翻譯--電動下肢輔助外骨骼的開發(fā)與分析【中文4480字】【PDF+中文WORD】,中文4480字,PDF+中文WORD,機械,外文,文獻,翻譯,電動,下肢,輔助,骨骼,開發(fā),分析,中文,4480,PDF,WORD Journal of Bionic Engineering 14 (2017) 272–283 Development and Analysis of an Electrically Actuated Lower Extremity Assistive Exoskeleton Yi Long1, Zhijiang Du1, Chaofeng Chen1, Weidong Wang1, Long He2, Xiwang Mao2, Guoqiang Xu2, Guangyu Zhao2, Xiaoqi Li2, Wei Dong1 1. State Key Laboratory of Robotics and System, Harbin Institute of Technology (HIT), Harbin 150001, China 2. Weapon Equipment Research Institute, China South Industries Group Corporation, Beijing 102202, China Abstract An electrically actuated lower extremity exoskeleton is developed, in which only the knee joint is actuated actively while other joints linked by elastic elements are actuated passively. This paper describes the critical design criteria and presents the process of design and calculation of the actuation system. A flexible physical Human-Robot-Interaction (pHRI) measurement device is designed and applied to detect the human movement, which comprises two force sensors and two gasbags attached to the inner surface of the connection cuff. An online adaptive pHRI minimization control strategy is proposed and implemented to drive the robotic exoskeleton system to follow the motion trajectory of human limb. The measured pHRI information is fused by the Variance Weighted Average (VWA) method. The Mean Square Values (MSV) of pHRI and control torque are utilized to evaluate the performance of the exoskeleton. To improve the comfort level and reduce energy consumption, the gravity compensation is taken into consideration when the control law is designed. Finally, practical experiments are performed on healthy users. Experimental results show that the proposed system can assist people to walk and the outlined control strategy is valid and effective. Keywords: exoskeleton, pHRI measurement, data fusion, pHRI minimization, adaptive control 1 Introduction A lower extremity exoskeleton is an automatic device worn by human users to improve strength and en-durance. In recent decades, many advances and progress have been made in the development of wearable exoskeletons. The Berkeley Lower Extremity Exoskeleton (BLEEX) was designed to assist people to carry heavy loads, which could walk at the speed of 0.9 m·s–1 while carrying 34 kg payload[1]. A single robot leg has four actuators which leads to the complexity and heaviness of the whole system[2]. The developed exoskeletons latter, i.e., ExoHiker, ExoClimber and HULC, simplify the mechanical structure and reduce the number of active Degrees of Freedoms (DoFs). Those DoFs of joints with the highest power consumption during gait cycles should be actuated while the rest DoFs are passively connected with elastic elements to simplify the system[3]. Generally, the lower extremity exoskeleton leg is composed of serial or serial-parallel leg segment. A cam-mechanism is implemented in the design of ankle joint of the exoskeleton[4]. A novel serial-parallel mechanism is designed in a lower extremity exoskeleton to augment load-carrying for walking[5,6]. However, parallel mechanism will increase the complexity of mechanism and decrease its portability. Compact serial-link mechanisms are utilized in many advanced exoskeletons, e.g., Ekso[7] and ReWalk[8]. The actuation system design and development is a significant aspect for the exoskeleton. There are several popular actuation system modes that have been applied in lower extremity exoskeleton, including hydraulic actuators, electrical motors, serial elastic actuators and pneumatic muscle actuators. Torque-mass ratio, velocity, range of motion and control ability are significant specifications when choosing actuators for exoskeletons[9]. Hydraulic actuators and electrical motors are utilized frequently due to their high torque-mass rates[10].Comprehensively, electrical motors are suitable for exoskeletons due to its wide range of motion, ease of control and maintenance, and simplicity of the system. Normally, the electrical motors are placed parallel with the joint of mechanical legs, which causes increment of structure complexity. Different from the traditional electrical actuators, a novel electrical actuator consisting of DC motor, gear pair and ball screw is developed, which has a more compact mechanical structure com-pared with the traditional actuator. pHRI-based human motion intent estimation is a critical step for the exoskeleton control. It is essential to get an accurate measurement of the pHRI for the robotic exoskeleton control and the assessment of the assistance grade[11]. The connection cuff is a widely-used device to fasten the human limb to the exoskeleton, which affects the wearing comfort and the walking performance. A flexible connection cuff is a suitable device to interact with the human user. A flexible sensor technique is developed to measure the pHRI pressure, where the sensory system is composed of several optical-electronic sensors[12]. The system has at least six sensors on the surface of the cuff. The pHRI force can be measured by a strain gauge, where a circle sensor is utilized[13]. How-ever, the structure of the interaction cuff is complex and has multiple sensors. In this work, a flexible pHRI measurement device is designed and applied in the robotic exoskeleton control, which is composed of two gasbags and one force sensor connected to each gasbag. The gasbag can enlarge the interaction area and guarantee that the pHRI can be measured easily. In addition, the usage of gasbag can increase the wearing comfort and adapt to different human users. The control strategy design is the core issue for the exoskeleton. Control strategies can be divided into three categories according to methods of estimating human motion intent, i.e., approaches based on signals measured from the human body, approaches based on inter-action force measurement and approaches based on signals measured from exoskeletons[14]. In order to guarantee the natural gait of human users, sensitivity amplification control, model-based control and hybrid assistive strategy are suitable for load-carrying task[15].However, sensitivity amplification control is sensitive to external disturbances and model-based control is dependent on the accurate dynamic model.Compared with other assistive strategies, hybrid control strategy is applied successfully in exoskeletons for load-carrying and shows an increased performance. The pHRI-based control strategy is an effective approach for exoskeleton. To minimize the pHRI force, a RBF neural network was proposed to compensate the dynamic uncertainties, where there is no force sensor in the interaction cuff[16]. However, there are some sensors on the human limbs. The goal of human exoskeleton collaboration is to reduce or eliminate the misalignment between the human user and the exoskeleton. In this work, we propose an online adaptive strategy to drive the robotic exoskeleton. With the discussion above, we highlight contributions of this paper: 1) A lightweight and compact electrically actuated robotic exoskeleton system is designed and implemented for walking assistance. 2) A pHRI measurement device is designed, which consists of a force sensor and a gasbag embedded in the connection cuff.3)A control strategy called model-free adaptive pHRI minimization is proposed to drive the exoskeleton to follow the human limbs. The remainder of this paper is organized as follows. The specific mechanical design of exoskeleton and the details of actuator are given in the second section. The sensory system is illustrated in the third section. In the fourth section, the adaptive pHRI minimization control strategy is proposed. Experiments are performed and experimental results are presented in the fifth section. Discussion and conclusion are presented in the final section. 2 System architecture description 2.1 Mechanical structure Exoskeletons are anthropomorphic devices that work in parallel with the human body. In the design of lower extremity exoskeletons, the number of DoF is required to be close to the number of human lower limbs’ DoFs in order to achieve a comfortable walking[17]. The lower extremity assistive exoskeleton has two mechanical segments attached to human limbs, in which there are seven DoFs on the thigh, the knee and the ankle joints in total. Fig. 1 shows the architecture of the lower extremity exoskeleton, which includes three main components as below[18]: (1)The leg segment including a shank and a thigh is attached to the human user leg by the interaction cuffs. The knee joint is activated by an actuator while the hip joint is passive. The actuator of the knee joint consists of a DC motor and a ballscrew, which helps legs walk and supports the weight of system when the leg is in the stance phase. The leg segment is made of aluminum and its weight is about 4 kg. The leg segment is attached to the waist part of the exoskeleton through a connection mechanism. All wires of the control module are integrated into the mechanism through the wiring slot connecting to the central controller. The lower extremity exoskeleton is suitable for the user whose height is in the range of 168 cm – 188 cm, where the length of the robotic thigh can be adjusted in the range of 430 mm – 470 mm while the robotic shank can be adjusted in the range of 470 mm – 510 mm. (2)The trunk, which is used as a platform to integrate the control system, the power supplies and the payload to be carried. The trunk of the exoskeleton consists of a connector to the leg, an adjustment mechanism for the waist width, a backpack and an elastic element connecting the backpack with the robotic waist. The elastic element with enough stiffness is utilized to support the weight of the mechanism. The length of the robotic waist can be adjusted in the range of 340 mm – 360 mm. The backpack will be connected to an ergonomic mechanism which is tied to human torso to carry the control enclosure, the power unit and other equipments. (3)The wearable shoes connected with the shanks, each of which contains three pressure sensors to collect ground reaction force. The wearable shoe is made of rubber to enhance the flexibility. The thin upper sole is fixed by bolts. Three pressure sensors or foot switches are placed into the lower sole and all wires are integrated together. The wearable shoe is connected with the leg segment through the ankle joint, which is designed as a curved ball bearing with three DoFs. To make the weight with a flexible support, some elastic elements, e.g., flexible rubber plates, are placed into the connecting part between the ankle joint and the shoe. 3 Sensor system The sensor system is the outermost layer of the control prototype and consists of three kinds of sensors, i.e., the digital and analogue pressure sensor embedded in the insole, the force sensor with gasbag for pHRI measurement and the encoder to obtain the joint angular position. 3.1 Pressure sensor in the wearable shoe Two digital pressure sensors (foot switches) are placed in the wearable shoe to measure the interaction force between the exoskeleton system and the ground. As Fig. 4 shows, there is one sensor placed on the connection area to measure the weight of the exoskeleton system. The selected digital sensor outputs binary values, i.e., 0 and 1, which outputs 1 when the human user presses upon the sensor. The pressure sensor used in the connection area, which can output the voltage signal proportional to the applied force. This kind of sensor has enough resolution to record the GRF changes during the stance period. In addition, this kind of sensor is small and thin enough to be embedded into the wearable shoes and to ensure the wearing comfort. The used pressure sensor has a sensing area of 4.52 cm2 and has a measuring range from 0 kg to 100 kg. 3.2 Force sensor with gasbag Physical HRI measurement is a critical issue for exoskeleton control. Based on pHRI, the human motion intent can be obtained. The measurement range of the force sensor is from 0 kPa to 1 kPa Wearable shoe Gasbag based sensor Optical encoder and the hose has a diameter of 7.5 mm. The air pipeline is connected to the gasbag, the force sensor changes as the air pressure changes and the signal output wire is connected to the collection part. The force sensor signals can be collected and the range of output voltage is 0 V – 10 V and the sensitivity is 2.0 mv·V 1. The force sensor is convenient to measure the interaction signals between the human limb and the mechanical limb. The two gasbags are attached to the interaction cuff, each of which has an area of 100 mm by 70 mm. 3.3 Encoder on the mechanical joint The angular position of the exoskeleton joint is collected by the optical encoder. This sensor is a kind of incremental encoder (HEDSS, German), whose resolution is 1024 line for a rotation. The angular position is collected to illustrate the current mechanical position. The value of the optical encoder is recorded and transferred to the central controller. The specific parameters of all applied sensors are shown in Table 1. 4 Discussion and conclusion In this paper, the development and analysis of an electrically actuated lower extremity exoskeleton has been presented. Compared to previous wearable robots, the developed lower extremity exoskeleton has the fol-lowing features: (1)An integrated actuation system is proposed which consists of a DC motor, a gear transmission structure and a ballscrew. Compared with those robots whose actuators are placed directly in parallel with the mechanical joint, this kind of architecture is more compact and nice-looking. In the electrically actuated exoskeletons, the traditional structure in which the mo-tor is mounted directly in parallel with the joints greatly decreases the compactness of the whole system. In comparison, this kind of quadrangle-like placement can make the structure of the robot leg slim and compact. (2)A flexible pHRI measurement device is designed and implemented in the wearable exoskeleton. This kind of connection cuff can ensure the wearing comfort of the user. Compared with the HRI measurement devices, e.g., Refs. [12] and [13], the usage of gasbag is helpful to enlarge the pHRI measurement area and reduce the number of sensors. In this work, two gasbags are nearly filled in the connection cuff, and correspondingly two force sensors are connected with those two gasbags. (3)A model-free adaptive pHRI minimization methodology was proposed to drive the robotic exoskeleton to follow the human movement. The performance of the model-based control strategy for robotic exoskeleton, e.g., Ref. [1], is greatly dependent on the accuracy of the dynamic model of the system. The mode-based control strategy is complicated and needs a large number of complex practical tests to identify the robot model. The proposed method is simple and easy to apply. The control target of the robotic exoskeleton is to minimize the pHRI to achieve the human exoskeleton coordination. The pHRI can be regarded as the evaluation index to illustrate the wearing comfort. If the pHRI is smaller, the user will feel more comfortable. In addition, the usage of gravity compensation is capable of improving the wearing comfort and reducing energy consumption. The experiment wearing the exoskeleton is conducted on a healthy male and the experimental results show the proposed method can reduce the pHRI and control current over 95.7 ± 0.4% and 61.7 ± 6.4% respectively. The proposed control strategy is effective and valid for the lower extremity exoskeleton. Looking to-wards the future, we envision several improvements that would increase the performance of the developed lower extremity exoskeleton. One of the future works is to improve adaptability to different users. Acknowledgement The authors thank the anonymous reviewers for their constructive remarks and suggestions for improving this paper. References [1] Kazerooni H, Steger R, Huang L. Hybrid control of the Berkeley Lower Extremity Exoskeleton (BLEEX). Interna-tional Journal of Robotics Research, 2006, 25, 561–573. [2] Zoss A, Kazerooni H, Chu A. On the mechanical design of the Berkeley Lower Extremity Exoskeleton (BLEEX). Pro-ceeding of IEEE/RSJ International Conference on Intelli-gent Robots and Systems, Alberta, Canada, 2005, 3465–3472. [3] Miao Y, Gao F, Pan D. Mechanical design of a hybrid leg exoskeleton to augment load-carrying for walking. Interna-tional Journal of Advanced Robotic Systems, 2013, 10, 502–512. [4] Copilusi C, Ceccarelli M, Carbone G. Design and numerical characterization of a new leg exoskeleton for motion assis-tance. Robotica, 2014, 33, 1–16. [5] Miao Y, Gao F, Pan D. State classification and motion de-scription for the lower extremity exoskeleton SJTU-EX. Journal of Bionic Engineering, 2014, 11, 249–258. [6] Miao Y, Gao F, Pan D. Compound pendulum modeling and resonant frequency analysis of the lower limbs for the wearer and exoskeleton. Journal of Bionic Engineering, 2015, 12, 372–381. [7] Joanne P. The Pransky interview: Russ Angold, Co-founder and president of Ekso? labs. Industrial Robot, 2014, 41, 329–334. [8] Talaty M, Esquenazi A, Briceno J E. Differentiating ability in users of the ReWalk (TM) powered exoskeleton: An analysis of walking kinematics. Proceedings of IEEE In-ternational Conference on Rehabilitation Robotics, Seattle, USA, 2013, 1–5. [9] Onen U, Botsali F M, Kalyoncu M, Tinkir M, Yilmaz N, Sahin Y. Design and actuator selection of a lower extremity exoskeleton. IEEE/ASME Transactions on Mechatronics,2014, 19, 623–632. [10] Hollerbach J M, Hunter I W, Ballantyne J. A comparative analysis of actuator technologies for robotics. Robotics Re-view, 1991, 2, 299–342. [11] Santis A D, Siciliano B, Luca A D, Bicchi A. An atlas of physical human–robot interaction. Mechanism & Machine Theory, 2008, 43, 253–270. [12] Donati M, Vitiello N, De Rossi S M M, Lenzi T, Crea S, Persichetti A, Carrozza M C. A flexible sensor technology for the distributed measurement of interaction pressure. Sensors, 2013, 13, 1021–1045. [13] Del-Ama A J, Moreno J C, Gil-Agudo A, De-los-Reyes A, Pons J L. Online assessment of human-robot interaction for hybrid control of walking. Sensors, 2012, 12, 215–225. [14] Huo W, Mohammed S, Moreno J C, Amirat Y. Lower limb wearable robots for assistance and rehabilitation: A state of the art. IEEE Systems Journal, 2014, 10, 1068–1081. [15] Anam K, Al-Jumaily A A. Active exoskeleton control sys-tems: State of the art. Procedia Engineering, 2012, 41, 988–994. 電動下肢輔助外骨骼的開發(fā)與分析 摘要 現(xiàn)已研發(fā)出一種電機驅動下肢外骨骼機器人，其中只有膝關節(jié)主動運動，被彈性元件連接的其他關節(jié)則被動地驅動。本文介紹了關鍵的設計標準，也介紹了驅動系統(tǒng)的設計和計算過程。設計并應用了一種靈活的物理人機交互(pHRI)測量裝置，用于檢測人體運動，它包括兩個力傳感器和連接在連接袖口內表面的兩個氣囊。提出并實現(xiàn)了一種在線自適應pHRI最小化控制策略，以驅動機器人外骨骼系統(tǒng)跟蹤人類肢體的運動軌跡。通過方差加權平均值(VWA)法將測量的pHRI信息融合在一起。利用pHRI和控制扭矩的均方值（MSV）來評估外骨骼的性能。為了提高舒適性和降低能耗，設計控制律時會考慮重力補償。最后，對健康的用戶進行了實驗。實驗結果表明，該系統(tǒng)能有效地幫助人們行走，并且概述的控制策略是有效和有效的。 關鍵詞：外骨骼，pHRI測量，數(shù)據(jù)融合，pHRI最小化，自適應控制 1. 介紹 下肢外骨骼機器人是人類用戶為提高力量和耐力而穿戴的裝置。近幾十年來，可穿戴外骨骼機器人的發(fā)展取得了很大進步。伯克利下肢外骨骼（BLEEX）的設計目的是幫助人們承受重負荷，在承載34公斤有效載荷時可以以0.9 m·s-1的速度行走[1]。機器人的一條腿有四個驅動器，導致整個系統(tǒng)的復雜性和沉重[2]。后者開發(fā)的外骨骼，即ExoHiker，ExoClimber和HULC，簡化了機械結構并減少了活動自由度（DoFs）的數(shù)量。那些在步態(tài)周期中功耗最高的關節(jié)DoFs應該被激活，而其余DoFs被動連接彈性元件以簡化系統(tǒng)[3]。通常，下肢外骨骼腿由串行或串行平行的腿段組成。在外骨骼的踝關節(jié)設計中實現(xiàn)了一個凸輪機構[4]。一種新型的串并聯(lián)機構被設計在下肢外骨骼中以增加步行時的負重[5,6]。然而，并行機制會增加機制的復雜性并降低其可移植性。緊湊型串行鏈接機制被用于許多先進的外骨骼，例如Ekso [7]和ReWalk [8]。 制動系統(tǒng)的設計和開發(fā)是外骨骼的一個重要方面。有幾種流行的驅動系統(tǒng)模式已應用于下肢外骨骼，包括液壓執(zhí)行器、電機、串行彈性執(zhí)行器和氣動肌肉執(zhí)行器。在選擇外骨骼執(zhí)行器時，力矩質量比、速度、運動范圍和控制能力都是很重要的[9]。液壓執(zhí)行器和電動馬達由于其高轉矩質量率而被頻繁使用[10]。通常，電動馬達適用于外骨骼，因為其運動范圍廣泛，易于控制和維護，并且系統(tǒng)簡單。通常情況下，電動機與機械腿的接頭平行放置，導致結構復雜度增加。與傳統(tǒng)的電動執(zhí)行機構不同，開發(fā)了一種由直流電機，齒輪副和滾珠絲杠組成的新型電動執(zhí)行機構，與傳統(tǒng)的執(zhí)行機構相比，它具有更緊湊的機械結構。 基于pHRI的人體運動意圖估計是外骨骼控制的關鍵步驟。準確測量機器人外骨骼控制的pHRI并評估輔助等級是至關重要的[11]。連接袖口是將人的肢體緊固到外骨骼的廣泛使用的裝置，其影響穿著舒適性和行走表現(xiàn)。靈活的連接袖口是與人類用戶進行交互的合適設備。靈活的傳感器技術被開發(fā)用于測量pHRI壓力，感應系統(tǒng)由多個光電傳感器組成[12]。該系統(tǒng)在袖帶表面至少有六個傳感器。 pHRI力可以通過應變計測量，其中使用圓形傳感器[13]。然而，相互作用袖帶的結構非常復雜并且有多個傳感器。在這項工作中，靈活的pHRI測量裝置被設計和應用在機